基于RBF神经网络的一类非线性系统反演鲁棒自适应控制

Robust Adaptive Control for a Class of Nonlinear Systems Using Backstepping Based on RBF Neural Network

摘要:

提出一种基于RBF神经网络的一类非线性系统反演鲁棒自适应控制器设计方法。使用RBF神经网络逼近系统不确定性,并和控制器与虚拟控制器中的鲁棒项一起消除不确定性的影响,由Lyapunov稳定性理论推出的RBF神经网络权值矩阵的自适应律能保证闭环系统的所有信号有界,且误差能够全局指数收敛于原点的邻域。该方法不需要系统不确定性的上界以及其任意阶导数,最后的仿真结果验证了方法的有效性。

Abstract:

A robust adaptive controller design method using backstepping techniques based on RBF neural computing for a class of uncertain nonlinear systems was proposed. RBF neural networks were used to approximate the uncertainties and to eliminate the bad effects of the uncertainties with robust terms in the controller and virtual controllers. The adaptive tuning rules of RBF neural network weight matrixes were derived by the Lyapunov stability theorem that guaranteed all signals of the closed-loop system were bounded and error signals exponentially converged to a neighborhood of the origin globally. The proposed method did not need the supper bounds of the uncertainties and their any order derivatives. In the end, simulation results were presented to demonstrate the effectiveness of the proposed method.